Advance Access publication July 8, 2005.
3122© The Author 2005. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
Lack of seasonal variations in fertilization, pregnancy and
implantation rates in women undergoing IVF
D.M.Wunder
1,3, C.Limoni
2and M.H.Birkhäuser
1and the Swiss FIVNAT-Group
1Department of Obstetrics and Gynaecology, University of Berne, Berne and 2Department of Economics and Social Sciences (DSAS), University of Applied Sciences of Southern Switzerland (SUPSI), Manno, Switzerland
3To whom correspondence should be addressed at: Universitäts-Frauenklinik, Effingerstrasse 102, 3010 Bern, Switzerland. E-mail: dorothea.wunder@insel.ch
BACKGROUND: Several studies have investigated seasonal variations during IVF. Their results are contradictory, especially concerning fertilization and pregnancy rates. The aim of the present study was to re-evaluate these param-eters using a large number of IVF cycles. METHODS: A total of 7368 IVF cycles conducted in Switzerland between 1995 and 2003 were retrospectively analysed. To avoid a bias in the evaluation of the fertilization rate, only IVF cycles without ICSI were considered for analysis. Cycles were assigned to seasons according to the date of the begin-ning of stimulation. RESULTS: There were no statistically significant differences between the seasons concerbegin-ning the fertilization, the pregnancy and the implantation rates. However, statistically significant variables deciding on the outcome of an IVF cycle are age, centre, aetiology of infertility and day of transfer. CONCLUSIONS: There were no statistically significant seasonal differences in central Europe (Switzerland) that influenced the outcome of IVF treat-ment. The only statistically significant variables of IVF outcome were age, centre, aetiology of infertility and day of transfer. A change to routine fertility treatment concerning the different seasons should therefore not be taken into account.
Key words: fertilization rates/implantation rates/IVF/pregnancy rates/seasonality
Introduction
Every physician and biologist working in the field of reproduc-tive medicine keeps wondering about variable pregnancy rates, because success rates are fluctuating without visible reason from excellent to very poor. This is often not explained by the group a patient belongs to (e.g. advanced age of the woman, poor responders, primary infertility of the female, multiple infertility factors) nor by cultural conditions, technical failures or other comprehensible reasons. Thus, the question arises whether the different seasons of the year can have good or bad influences on the characteristics of the oocytes and on the ferti-lization and pregnancy rates during assisted reproductive tech-nologies in humans. If this were the case, it would be possible to avoid IVF treatment cycles in the ‘bad months’ in order to enhance pregnancy rates.
All over the world, seasonal changes in the reproductive life of animals are well known. The question arises whether human reproduction is also affected by such seasonal differences, although sexual activity of human beings is not bound to sea-sons. Several epidemiological studies point to a variation in natural conception and birth rates not only in animals but also in humans (Odegard et al., 1977; Mathers and Harris, 1983; Roenneberg et al., 1990a,b; Lam et al., 1991), being reduced during the spring in warm climates in non-equatorial regions (Huntington, 1938; Lamar et al., 1943; Rosenberg, 1966;
Becker, 1981). The reasons for this phenomenon are not very well understood; possible explanations might be the variations in semen quality, showing a significantly lower sperm count in the summer compared to the winter (Tjoa et al., 1982; Levine et al., 1988; Reinberg et al., 1988; Politoff et al., 1989; Saint Pol et al., 1989; Levine et al., 1990; Levine, 1991), variations in the ovulation rate (Timonen et al., 1964; Kivela et al., 1988; Rameshkumar et al., 1992) and variations in endometrial func-tion (Timonen et al., 1964; Kottler et al., 1989) throughout the seasons: these variations might be linked to the light–dark effect of the female reproductive axis (Rojansky et al., 1992). On the other hand, inconsistent frequencies of intercourse (Odegard et al., 1977; Ehrenkranz et al., 1983; Abas and Murphy, 1987; Jacobsen et al., 1987), possibly increasing dur-ing holidays (Rosenberg, 1966; Wrigley and Schofield, 1981; Cesario, 2002), might also be an explanation. Others, however, did not confirm the occurrence of varying frequencies of inter-course as a function of season (Udry et al., 1967; Levine et al., 1990).
Studies on reproductive outcome of assisted technologies in primates have shown significant variations throughout the sea-sons, the oocyte maturation being much poorer in winter months (Smith et al., 1978; Chan et al., 1982). The most intrigu-ing findintrigu-ing in these above-mentioned studies is the persistintrigu-ing seasonal effect on the maturation of retrieved oocytes during
IVF, despite the administration of exogenous gonadotrophins and constant environmental conditions.
Possible seasonal variations during assisted reproductive treatment of humans have been suggested by several studies (Wood et al., 1985; Stolwijk et al., 1994; Chamoun et al., 1995; Ossenbühn, 1998; Rojanski et al., 2000; Weigert et al., 2001), although variations of the ovulation rate due to varying endogenous gonadotrophin secretion are suppressed by the hormonal therapy given during IVF treatment. Furthermore, lower sperm counts in non-treated ejaculates (observed during the summer months) are compensated in IVF treatment cycles by utilizing a concentrated and constant amount of motile sperm to inseminate the oocytes. Other studies could not con-firm the presence of seasonal variations (Casper et al., 1988; Daya et al.,1993; Fleming et al., 1994).
Considering these controversial data, we retrospectively evaluated the outcome of 7368 IVF cycles, conducted over a 9 year period throughout Switzerland and encompassing all four seasons of the year.
Materials and methods
Patients
A total of 7368 IVF cycles, conducted in all 18 IVF clinics of Switzerland between 1995 and 2003, were retrospectively analysed (4688 patients, 62.3% with one cycle, 25.5% with two, 7.1% with three, 5.1% with more than three cycles). For this analysis, the national IVF data regis-ter (FIVNAT) has been used.
To avoid a bias in the evaluation of the fertilization rate, only IVF cycles were considered for analysis. No cycles with ICSI were included. The indications for the IVF treatment were tubal factors in 34.6%, male factor in 11.7%, idiopathic infertility in 9.5%, endometri-osis in 5.1% and multiple causes in 29.7%, while other causes accounted for 9.4%. Patients had in 60.7% of the cycles a primary infertility and in 39.3% a secondary infertility.
The decision for IVF treatment had been made before the start of the study and all patients have been treated routinely. Because all data have been rendered anonymous when entered in the computerized data base, approval of the ethical committee was not necessary. Methods
The seasons were defined before analysing the data, according to the calendar definitions of the seasons for Europe, each season lasting 3 months: Spring: March 21–June 20; Summer: June 21–September 20; Autumn: September 21–December 20; Winter: December 21–March 20. Because the season may not only have an impact on embryo devel-opment and implantation, but also influence the oocyte during its development at the time of IVF treatment, the attribution of the patients to the corresponding seasons was made according to the date of the beginning of stimulation.
Complete infertility investigation of both partners (hormonal evalu-ation, gynaecological ultrasound, hysterosalpingography or laparos-copy to evaluate tubal patency and semen analysis) preceded the IVF treatment.
In all, 68.2% of patients were stimulated with the long protocol, 32.8% with the short protocol. Down-regulation was achieved with triptorelin (daily injection or depot preparation; Decapeptyl®, Ferring
pharmaceuticals, Wallisellen, Switzerland), goserelin depot prepara-tion (Zoladex®, Astra Zeneca, Zug, Switzerland) or nafarelin nasal
spray (Synrelina®, Pharmacia, Dübendorf, Switzerland). Gonado-trophins used for stimulation were in 24.3% recombinant FSH
(Gonal-F®, Serono Pharma, Geneva, Switzerland; or Puregon®, Orga-non Pharmaceuticals, Pfäffikon, Switzerland), in 34.2% highly puri-fied FSH (Metrodin HP®, Serono; Fostimon®, IBSA, Pambio-Noranco, Switzerland) and 40.4% HMG, (Pergonal®, Serono; Hum-egon®, Organon; Menogon®, Ferring pharmaceuticals; Merional®, IBSA). The different stimulation protocols and gonadotrophins were statistically equally distributed among the different groups of patients divided according to the seasons, thus excluding any bias.
Thirty-five to 36 h after hCG administration, oocytes were retrieved by needle aspiration, with transvaginal ultrasound guidance and under routine intravenous sedation or general anaesthesia.
Cycles with micromanipulation (e.g. ICSI, subzonal injection of sperm etc.), with gamete or zygote intra-Fallopian transfer as well as cryocycles were excluded. The laboratory work (preparation of oocytes and the ejaculate, fertilization, incubation and embryo trans-fer) was performed by the use of previously described standard tech-niques (Lewin et al., 1986). Fertilization was assessed 17 h (range 15–20) after IVF. Only normal pronucleids [normally fertilized oocytes with two pronuclei (PN) and two polar bodies] were considered for embryo transfer and freezing was done if more than two or three pronucleids were available. Those with the highest PN score and with the best morphological grade were selected for transfer in each treatment cycle. They were cultured for another 20–30 h at 37°C in fresh CO2 -equilibrated IVF medium. All remaining pronucleids were cryopre-served in the 4-cell embryo or PN stage, for later transfers. Since the beginning of the year 2001, no embryos, but only pronucleids, are allowed to be frozen according to Swiss law.
The pregnancy rate after IVF is defined by the proportion of patients with a positive HCG value, ≥14 days after embryo transfer. The fertilization rate is defined by the proportion of only mature oocytes with one polar body resulting in pronucleids. The implanta-tion rate is defined by the proporimplanta-tion of the number of embryos trans-ferred resulting in gestational sacs ultrasonographically diagnosed 2 weeks after the positive pregnancy test, including ectopic gestations. Mean embryo quality could not be calculated, because the different IVF centres used different classifications for the evaluation of the embryo quality, making statistical analysis impossible.
Statistical methods
Categorical variables were compared for homogeneity within sea-sons and months using χ2 for goodness of fit. Continuous variables were compared using one-way parametric analysis of variance and multivariate regression. IVF results of the cycles in the different sea-sons and months were assessed first using χ2 goodness of fit and
subsequently using a multivariate logistic model adjusting for cen-tre, age, indication, infertility type and year of stimulation. In a sub-sequent analysis considering only cycles with embryo transfer, two additional covariates—the number of transferred embryos and the day of transfer—were considered. The significance level was set to
α = 0.05 two-tailed and all calculations were performed using SAS vs 8.02.
All statistical analyses were repeated considering the first treatment cycle only.
Results
A total of 7368 IVF cycles (first treatment cycle in 62.2%, second cycle in 25.6%, ≥ third cycle in 12.2%) were analysed. To exclude any potential bias due to repeated observations, a subsequent statistical sub-analysis was carried out for the 4688 IVF first cycles. The results from the sub-analysis performed with the 4688 first cycles did not differ from those obtained by
the analysis of all the cycles (data not shown). Therefore, all IVF cycles could be evaluated.
Table I shows the distribution of the clinical parameters of the patients (age, primary/secondary infertility, aetiology of infertility and number of IVF cycles) in the different seasons. There were no significant differences for the type of infertility (primary/secondary) within the seasons, but the aetiology of infertility (P < 0.01) and the number of IVF cycles conducted within the different seasons varied significantly. The number of IVF cycles performed was lowest in summer and autumn and highest in winter and spring (P < 0.0001).
Table II presents the results of the number of IVF cycles, the number of oocytes retrieved, the number of pronucleids obtained, the number of embryos transferred, the pregnancy rates and the implantation rates for the IVF cycles done in the different seasons. Pregnancy rates were not found to be different between the seasons. Table III presents the results of the pregnancy and implanta-tion rates for IVF cycles per month during the whole observa-tion period. There were no significant differences in pregnancy and implantation rates, but there was in the number of embryos transferred. Also after correction for the number of embryos, age, infertility type (primary/secondary), aetiology of
infertil-ity and day of transfer, there were no significant differences in adjusted pregnancy and implantation rates (data not shown).
Table IV shows the distribution of the number of IVF cycles, the age of the patients, the type infertility (primary/secondary), the aetiology of infertility, the number of oocytes retrieved, the number of pronucleids obtained and the number of embryos trans-ferred in the different years of the observation period. For all these parameters—with the exception of the type of infertility—signi-ficant differences have been seen between years of stimulation.
Table V presents the pregnancy results of all initiated IVF cycles adjusted for season, age of the patient, infertility type, treatment indications, year of treatment and centre. The odds of becoming pregnant do not differ between the season (OR 0.98; CI: 0.83–1.17 for summer, OR 1.02; CI: 0.85–1.21 for autumn and OR 1.11; CI: 0.94–1.31 for winter, taking spring as reference), whereas age, indication for treatment, year of stimulation and centre are significant factors: the odds of a patient aged <30 years are 4.4-fold those of a patient aged ≥40 years (OR CI: 3.6–5.7), for patients aged 30–34 years they are 3.4-fold (OR CI: 2.7– 4.3) and for patients aged 35 years 2.8-fold (OR CI: 2.3–3.6). The odds for the indication for treatment are unfavourable for a male factor (OR 0.74; CI 0.60–0.91).
Table I. Distribution of the age of the patients, infertility type (primary/secondary), aetiology of infertility (treatment indications) and number
of IVF cycles in the different seasons.
Values are percentages unless otherwise specified.
P-Values were obtained by one-way ANOVA for continuous variables and goodness of fit χ2 test for categorical variables. NS = not significant.
Clinical parameter Spring Summer Autumn Winter P
Age of patients (mean ± SD) 35.5 ± 4.4 35.5 ± 4.2 35.0 ± 4.5 35.6 ± 4.4 <0.0001
Infertility type Primary 59.3 61.2 62.2 60.3 NS Secondary 40.7 38.8 37.8 39.7 NS Treatment indications Multiple causes 30.5 29.8 28.1 30.3 NS Tubal 36.1 34.1 34.6 33.6 NS Male 10.9 10.3 12.4 12.8 NS Endometriosis 5.6 5.2 4.6 5.1 NS Idiopathic 7.9 10.6 10.8 8.9 0.01
Others (not specified) 9.0 10.1 9.4 9.3 NS
Initiated IVF cycles [n (%)] 1895 (25.7) 1667 (22.6) 1764 (23.9) 2042 (27.7) <0.0001
Table II. Number of initiated cycles, number of oocytes retrieved, number of pronucleids (normally fertilized oocytes with two pronuclei and two polar bodies)
obtained, number of embryos transferred, pregnancy rates and implantation rates for IVF cycles in the different seasons
Values given are mean and range; P values for continuous variables were obtained by Student’s t-test and for categorical variables with χ2 for goodness of fit. NS = not significant.
Clinical parameter Spring Summer Autumn Winter P
No. of initiated IVF cycles 1895 1667 1764 2042 <0.0001
Cycles with oocyte retrieval (% of initiated cycles) 1730 (91.3) 1539 (92.3) 1622 (92.0) 1864 (91.3) <0.0001 Cycles with transfer (% of initiated oocyte retrieval cycles) 1544 (81.5) 1343 (80.6) 1435 (81.3) 1693 (82.9) <0.0001
Oocytes retrieved (mean ± SD) 8.8 ± 6.1 8.5 ± 5.9 9.2 ± 6.6 8.5 ± 5.8 NS
Pronucleids obtained (mean ± SD) 5.2 ± 4.3 4.9 ± 4.2 5.1 ± 4.4 5.0 ± 4.1 NS
Embryos transferred (mean ±SD) 1.9 ± 1.1 1.8 ± 1.1 1.8 ± 1.0 1.9 ± 1.0 NS
Clinical pregnancies 361 322 353 411
Pregnancy rate by initiated cycle 19.1 19.3 20.0 20.1 NS
Pregnancy rate by oocyte retrieval 20.9 20.9 21.8 22.0 NS
Pregnancy rate by transfer cycle 23.4 24.0 24.6 24.3 NS
Ta
ble
III
.
Pr
egnancy and impla
n tation ra tes for all I V F cycl es b y mon th ov er the en tire ob serv atio n perio d NS = not significant. Clinic al paramete r Ja n Feb M a r Apr M ay June July A u g S ept O ct N o v D e c P Initiated cycles 715 7 37 714 6 37 766 414 3 71 655 737 7 31 669 222 < 0.0001 Cycle s with ooc y te retr ieval (% of ini tiated cycles) 9 1 .9 91 .6 91 .6 9 1 .2 90. 3 9 2 .5 9 1. 9 93. 0 9 0 .0 9 2. 9 93. 7 8 7 .4 < 0. 00 01 Cycle s with tr ansf er ( % of initiate d oocyte r etrie val cycles) 8 2 .4 82 .8 85 .4 8 1 .5 79. 0 8 2 .9 7 9. 2 81. 4 7 7 .9 8 3. 3 82. 7 7 9 .3 < 0. 00 01 Oocyte s retr ieved (mean ± SD) 8 .3 ± 5 .4 8 .8 ± 6. 0 8 .4 ± 5. 8 8 .5 ± 5 .7 8 .9 ± 6. 1 8 .7 ± 6. 3 9 .1 ± 6 .6 8 .5 ± 5. 6 8 .3 ± 6. 1 9 .3 ± 6 .8 9 .5 ± 6. 6 9 .3 ± 6. 8 N S Pr on ucleid s obtained ( m ean ± S D ) 4 .8 ± 3 .8 5 .2 ± 4. 2 5 .0 ± 4. 2 5 .1 ± 4 .3 5 .3 ± 4. 4 5 .0 ± 4. 2 5 .1 ± 4 .5 4 .7 ± 4. 0 4 .8 ± 4. 2 5 .2 ± 4 .4 5 .3 ± 4. 6 5 .5 ± 4. 4 N S Embr yos tr ans ferr ed (mean ± SD) 1 .9 ± 1 .0 1 .9 ± 1. 1 1 .9 ± 1. 0 1 .8 ± 1 .1 1 .8 ± 1. 1 1 .9 ± 1. 1 1 .8 ± 1 .1 1 .8 ± 1. 1 1 .7 ± 1. 1 1 .8 ± 1 .0 1 .9 ± 1. 0 1 .7 ± 1. 0 0 .01 Clin ical pr egnan c ies 1 48 15 7 1 4 5 1 0 9 143 82 6 8 125 14 7 1 40 128 55 Pregnanc y rate by initiated c y cle (%) 2 0.7 2 1. 3 2 0.3 1 7.1 18.7 1 9. 8 1 8.3 19.1 1 9. 9 1 9.2 19.1 2 4. 8 N S Pr egnan c y r ate b y oocy te r etrieval (%) 2 2. 5 2 3 .32 2 .2 1 8 .82 0 .72 1 .41 9 .92 0 .52 2 .22 0 .62 0 .42 8 .4N S Pr egnan c y r ate b y tran sfer cy cle (% ) 2 5. 1 2 5 .7 2 3 .8 2 1. 0 23. 6 2 3 .9 2 3. 1 23. 5 2 5 .6 2 3. 0 23. 1 3 1 .3 N S Im p lan tatio n rate (%) 1 4. 7 1 5 .1 1 3 .5 1 2. 0 12. 8 1 4 .1 1 2. 5 13. 6 1 3 .9 1 2. 9 11. 8 1 7 .8 N S
The year of stimulation is also a significant factor: being stimulated in 1997 (OR 0.6; CI 0.5–0.8) or 1998 (OR 0.7; CI 0.6–0.9) gives significantly worse odds for a pregnancy, as compared to 1995.
Table VI presents the pregnancy results limited to IVF cycles with embryo transfer adjusted for season, age of the patient, infertility type, treatment indication, number of trans-ferred embryos, day of transfer, year of treatment and centre.
The odds of becoming pregnant do not differ between the sea-sons (OR 0.94; CI: 0.79–1.13 for summer, OR 1.00; CI: 0.83– 1.20 for autumn and OR 1.09; CI: 0.91–1.30 for winter, taking spring as reference). In contrast, age, indication for treatment, number of transferred embryos, day of transfer, year of stimu-lation and centre are significant factors deciding on the out-come: the odds of a patient aged <30 years are 4.0-fold those of a patient aged ≥40 years (CI 3.0–5.3), for patients aged 30–34 Table IV. Distribution of the clinical parameters and number of IVF cycles in the different years of the observation period
NS = not significant.
Clinical parameter 1995 1996 1997 1998 1999 2000 2001 2002 2003 P
No. of IVF cycles 662 737 952 960 852 829 756 864 756 <0.0001
Age of patients (years) (mean ± SD) 34.8 ± 4.4 34.7 ± 4.4 34.7 ± 4.3 35.2 ± 4.3 35.5 ± 4.3 35.9 ± 4.4 35.9 ± 4.4 36.0 ± 4.4 35.9 ± 4.5 <0.0001 Type of infertility Primary 62.8 60.8 62.3 62.2 58.0 61.2 58.6 60.1 60.3 NS Secondary 37.2 39.2 37.7 37.8 42.0 38.8 41.4 39.9 39.7 NS Treatment indications Multiple causes 28.7 23.6 25.2 22.4 24.2 29.1 44.3 38.0 34.3 <0.0001 Tubal 28.9 35.1 33.5 38.1 39.1 36.8 31.6 33.3 33.2 0.015 Male 17.4 14.0 12.8 11.3 11.2 11.0 7.8 8.1 12.7 <0.0001 Idiopathic 5.0 4.7 4.8 5.2 5.8 5.1 2.4 6.4 6.7 0.003 Endometriosis 7.3 9.1 9.9 9.1 12.0 11.2 9.0 7.3 10.1 0.017
Others (not specified) 12.8 13.4 13.8 14.0 7.9 6.9 4.9 6.9 3.0 <0.0001
Oocytes retrieved (mean ± SD) 8.4 ± 6.0 8.4 ± 5.9 8.0 ± 5.4 8.4 ± 5.9 8.7 ± 6.0 8.9 ± 5.9 9.8 ± 7.1 9.1 ± 6.4 9.3 ± 6.1 <0.0001 Pronucleids obtained (mean ± SD) 5.1 ± 4.4 4.9 ± 4.2 4.5 ± 3.9 4.7 ± 4.0 4.8 ± 4.2 5.1 ± 4.3 5.7 ± 4.7 5.5 ± 4.4 5.4 ± 4.2 <0.0001 Embryos transferred (mean ± SD) 2.1 ± 1.2 2.0 ± 1.1 1.9 ± 1.1 1.8 ± 1.1 1.8 ± 1.1 1.9 ± 1.0 1.8 ± 1.0 1.7 ± 0.9 1.7 ± 0.9 <0.0001
Pregnancy rate by initiated cycle 19.8 18.9 15.0 17.1 19.1 20.5 19.4 24.7 23.4 <0.0001
Pregnancy rate by oocyte retrieval 21.1 20.5 16.1 18.6 20.7 22.5 22.0 27.2 25.7 <0.0001
Pregnancy rate by transfer cycle 24.0 23.1 18.3 21.3 24.0 24.7 23.9 30.0 28.5 <0.0001
Implantation rate 12.8 12.0 9.6 11.7 13.7 13.8 14.1 17.9 17.7 <0.0001
Table V. Pregnancy status for IVF-initiated cycles by season: multivariate odds ratios
aSignificant favourable/unfavourable factor, P < 0.05. bDetailed odds ratios by centre not shown.
NS = not significant.
Factors Point estimate 95% confidence interval
Lower Upper
Season (reference class: spring), P = 0.447 (NS)
Summer 0.981 (NS) 0.825 1.165
Autumn 1.015 (NS) 0.853 1.208
Winter 1.110 (NS) 0.939 1.313
Age (years) (reference class: ≥40 years), P < 0.0001
<30 4.375a 3.359 5.698
30–34 3.429a 2.713 4.333
35–39 2.840a 2.259 3.570
Indication (reference class: tubal), P = 0.0066
Endometriosis 0.959 (NS) 0.726 1.267
Idiopathic 1.207 (NS) 0.976 1.492
Male 0.737a 0.596 0.911
Others (not specified) 0.957 (NS) 0.829 1.104
Infertility type (reference class: primary), P = 0.0734 (NS)
Secondary 1.122 (NS) 0.989 1.272
Year (reference class: 1995), P = 0.0015
1996 0.888 (NS) 0.678 1.165 1997 0.628a 0.4801 0.823 1998 0.722a 0.554 0.941 1999 0.781 (NS) 0.597 1.023 2000 0.819 (NS) 0.625 1.072 2001 0.782 (NS) 0.592 1.032 2002 1.053 (NS) 0.806 1.376 2003 0.954 (NS) 0.738 1.285 Centre, bP < 0.0001
years they are 2.9-fold (CI 2.3–3.7) and for patients aged 35 years 2.5-fold (OR CI 2.0–3.2). The odds for the indication for treatment are more favourable to idiopathic (OR 1.3; CI 1.0– 1.6) when taking as reference the female indication.
The number of transferred embryos is also a significantly predictive factor: when taking as reference two transferred embryos the odds are significantly unfavourable for one embryo (OR 0.4; CI 0.3–0.5) and favourable for three or more embryos (OR 1.2; CI 1.0–1.4). The day of transfer is a signifi-cantly favourable factor when the transfer is done on day 4 or later, as compared to day 2 (OR 1.7; CI 1.2–2.5), while trans-fers performed on day 3 do not show a significant advantage compared to day 2 (OR 1.21; CI: 1.00–1.47). The year of stim-ulation is also a significant factor: being stimulated in 2002 (OR 1.7; CI 1.3–2.2) or 2003 (OR 1.5; CI 1.2–2.0) gives signif-icantly higher odds for pregnancy, as compared to 1995.
Discussion
Studying carefully the literature on seasonal variations in the out-come of IVF cycles, the extent of discrepancies between these studies is remarkable. One point is the difference in the stimula-tion protocols used. In some studies, cycles with down-regulastimula-tion were included, in others not. In addition, there was no agreement
between the studies about the criterion for the inclusion of a patient to the corresponding season: in some studies it was the day of the beginning of stimulation, in others it was the day of oocyte retrieval, resulting in a difference up to 17 days. Also, the pub-lished studies have been conducted in different climates. Because of the widely diverse environments with different temperatures and light changes, the different studies are not comparable. Regarding the heterogeneous study results, it is also remarkable that the seasonal effects observed are very complex. Rojanski et al. (2000) for example showed that even though the fertilization and the embryo quality rates were best in spring, the pregnancy rates were worst in spring. On the other hand, he found the worst fertilization and worst embryo quality rates in autumn, but the best pregnancy rates at that time (Rojanski et al., 2000).
The results between the different studies were highly inconsist-ent, some showing the best pregnancy rates in the months of November, December, January and February (Stolwijk et al., 1994), others showing the worst pregnancy rates in the months of January, February and March in IVF cycles with spontaneous LH surge (Casper et al., 1988), whereas no differences were seen in IVF cycles with HCG administration (Casper et al., 1988).
Our results showed no significant seasonal differences in the fertilization rate nor in the pregnancy or implantation rates. Our findings are in agreement with other studies also showing Table VI. Pregnancy status for IVF cycles with embryo transfer by season: multivariate odds ratios
aSignificant favourable/unfavourable factor, P < 0.05. bDetailed odds/ratios by centre not shown.
NS = not significant.
Factors Point estimate 95% confidence interval
Lower Upper
Season (reference class: spring), P = 0.4297 (NS)
Summer 0.943 (NS) 0.788 1.129
Autumn 0.999 (NS) 0.833 1.198
Winter 1.086 (NS) 0.912 1.294
Age (years) (reference class: ≥40 years), P < 0.0001
<30 3.979a 3.014 5.252
30–34 2.898a 2.272 3.696
35–39 2.496a 1.969 3.163
Indication (reference class: tubal), P = 0.0453
Endometriosis 1.051 (NS) 0.787 1.404
Idiopathic 1.290a 1.033 1.611
Male 0.852 (NS) 0.682 1.064
Others (not specified) 1.081 (NS) 0.932 1.255
Infertility type (reference class: primary), P = 0.4863 (NS)
Secondary 1.048 (NS) 0.919 1.193
No. of embryos transferred (reference class: 2), P < 0.0001
1 embryo 0.389a 0.301 0.503
≥3 embryos 1.175a 1.008 1.371
Day of transfer (reference class: day 2), P = 0.0119
Day 3 1.211 (NS) 0.997 1.470
Day 4 or later 1.687a 1.161 2.451
Year (reference class: 1995), P = 0.0107
1996 0.919 (NS) 0.693 1.217 1997 0.697a 0.526 0.924 1998 0.793 (NS) 0.598 1.052 1999 0.838 (NS) 0.627 1.120 2000 0.795 (NS) 0.593 1.065 2001 0.776 (NS) 0.573 1.051 2002 1.124 (NS) 0.835 1.513 2003 1.019 (NS) 0.749 1.385 Centre, bP < 0.0001
no differences in the results of assisted reproductive techniques done in the different seasons of the year (Casper et al., 1988; Daya et al., 1993; Fleming et al., 1994). One explanation why no differences in IVF success rates were found is that the lower sperm count observed during the summer months in non-treated ejaculates is compensated in IVF by utilizing a concen-trated and constant amount of motile sperm to inseminate the oocytes. Another explanation is the elimination of a seasonal influence on oocyte growth and ovulation by the suppression of the hypothalamic–pituitary axis through down-regulation by GnRH analogues and the exogenous administration of gonado-trophins for ovulation induction during the IVF cycle.
Our results also show no significant seasonal differences in the implantation rate, pointing to an absence of a seasonal change in endometrial receptivity.
Comparing our analysis of the different months, the best pregnancy and implantation rates were found in December, despite a significantly lower number of embryos transferred. However, these differences did not reach statistical signifi-cance. In a study conducted in Austria (Weigert et al., 2001), having roughly the same climate as Switzerland and having also included a large number of patients in a similar time-frame (8184 IVF cycles from 1992 to 1999), the best pregnancy rates were also found in December, and reaching statistical signifi-cance. Our results correspond only partly to the analysis of Stolwijk et al. (1994) from The Netherlands (quite a similar climate, too) who reported a better fertilization rate, embryo quality and pregnancy rate in the period from November to February. These European findings are difficult to explain. One hypothesis of seasonal fluctuations in fertilization and pregnancy rates is the correlation with the annual changes in sunlight exposition; hypothalamic–pituitary output, neuro-transmitters and melatonin are suspected to be causally related. The mechanism of the suspected seasonal variation in human fertility has been attributed to a direct melatonin or neurotrans-mitter effect on the end-organ (Kauppila et al., 1987; Yie et al., 1995; Malpaux et al., 1999), but the role of melatonin in repro-duction has been not well defined until now. However, influ-ences of the light/dark cycle on the female reproductive process have been demonstrated (Timonen et al., 1964; Kivela et al., 1988; Kottler et al., 1989; Rojansky et al., 1992), with a reduced ovulation rate and endometrial receptivity in winter months in spontaneous cycles. These findings are in frank con-tradiction to the cited studies on IVF cycles, where the best pregnancy rates were found in the winter. However, spontane-ous cycles cannot be compared with IVF cycles, because of the suppression of the hypothalamic–pituitary axis through down-regulation by GnRH analogues and the exogenous administra-tion of gonadotrophins for ovulaadministra-tion inducadministra-tion during in assisted reproduction treatment.
Our finding of the significantly different aetiologies between the seasons is difficult to explain and might be a coincidence. In contrast, the decrease of the male factor as an indication for IVF treatment from 1995 to 2003 is well explained by the increased use of ICSI for this aetiology since the mid-1990s.
Comparing the success rates in the different years of the observation period, a big increase in the quality of IVF treatment can be observed. The pregnancy rate increased significantly
even though the number of transferred embryos significantly decreased and the mean patient age significantly increased. There exist also statistically significant differences in the suc-cess rates between the different centres, but no single IVF cen-tre showed a seasonal variation or other statistically significant variations during the year in its results.
Our results showed that the statistically significant variables influencing the outcome of an IVF cycle are age, aetiology of infertility, day of transfer and centre. This concurs with the assisted reproductive technologies data report of 2002 (http:// www.cdc.gov/reproductivehealth/ART02/PDF/ART2002.pdf), conducted by the Centers for Disease Control.
In conclusion, by the analysis of a large number of IVF cycles, we confirmed that statistically significant variables for IVF outcome are age, aetiology of infertility, the day of trans-fer and centre. However, the suspected seasonal variability of the outcome of IVF cycles has not been confirmed: there is no statistically significant variability in fertilization, implantation or pregnancy rates between the seasons in IVF. A change of routine fertility treatments concerning the different seasons should therefore not be taken into consideration.
Acknowledgements
20 IVF centres (Baden; Basel (3); Bellinzona; Bern (2); Chene-Bougeries; Geneve; Kreuzlingen; Lausanne (2); Locarno; Lucern; Schaffhausen; St Gallen; Winterthur; Zurich (3)) of the FIVNAT-CH group participated in the study.
References
Abas M and Murphy D (1987) Seasonal affective disorders: the miseries of long dark nights? Br Med J 295,1504–1505.
Becker S (1981) Seasonality of fertility in Matlab, Bangladesh. J Biosoc Sci 13,97–105.
Casper RF, Erskine HJ, Armstrong DT, Brown SE, Daniel SA, Graves GR and Yuzpe AA (1988) In vitro fertilization: diurnal and seasonal variation in luteinizing hormone surge onset and pregnancy rate. Fertil Steril 49,644–648.
Cesario SK (2002) The “Christmas Effect” and other biometeorologic influ-ences on childearing and the health of women. J Obstet Gynecol Neonatal Nurs 31,526–535.
Chamoun D, Udoff L, Scott L, Madger L, Adashi EY and McClamrock HD (1995) A seasonal effect on pregnancy rates in an in vitro fertilization pro-gram. J Assist Reprod Genet 12,585–589.
Chan PJ, Hutz RJ and Dukelow WR (1982) Nonhuman primate in-vitro fertili-zation: seasonality, cumulus cells, cyclic mononucleotides, ribonucleic acid and viability assays. Fertil Steril 38,609–615.
Daya I, Garcia JE, Smith RD and Padilla SL (1993) Seasonal variation in in-vitro fertilization pregnancy rate: an analysis of 2,674 oocyte retrievals. Infertility 15,21–26.
Ehrenkranz JR (1983) Seasonal breeding in humans: birth records of the Labrador Eskimo. Fertil Steril 40,485–489.
Fleming C, Nice L, Hughes AO and Hull MGR (1994) Apparent lack of sea-sonal variation in implantation rates after in vitro fertilization. Hum Reprod 9,2164–2166.
Huntington E (1938) Season of birth: its relation to human abilities. John Wiley, New York.
Jacobsen F, Wehr R, Sack D and Rosenthal N (1987) Seasonal affective disor-ders, a review of the syndrome and its public health implications. Am J Publ Hlth 77,57–59.
Kauppila A, Kivela A, Parakrinen A and Vakkuri O (1987) Inverse seasonal rela-tionship between melatonin and ovarian activity in humans in a region with a strong seasonal contrast in luminosity. J Clin Endocrinol Metab 65,823–838. Kivela A, Kauppila A, Ylostalo P and Vakkuri O (1988) Seasonal, menstrual
and circadian secretions of melatonin, gonadotrophins and prolactin in women. Acta Physiol Scand 132,321–327.
Kottler ML, Coussieu C, Valensi P, Levy F and Dergrelle H (1989) Ultradian, circadian and seasonal variation of plasma progesterone and LH concentra-tions during the luteal phase. Chronobiol Int 6,267–277.
Lam DA and Miron JA (1991) Seasonality of birth in human population. Soc Biol 38,51–78.
Lamar JK and Rodgers R (1943) Season and human fertility in Galveston, Texas. Anat Rec 87,453–454.
Levine RJ, Bordson BL, Mathew RM, Brown MH, Stanley JM and Starr TB (1988) Deteriorations of semen quality during summer in New Orleans. Fertil Steril 49,900–907.
Levine RJ, Mathew RM, Chenault CB, Brown MH, Hurtt ME, Bentley KS, Mohr KL and Working TK (1990) Differences in the quality of semen in outdoor workers during summer and winter. New Engl J Med 323,12–16. Levine RJ (1991) Seasonal variation in human semen quality. In Zorginotti AW
(ed) Temperature and Environmental Effects on the Testis. Plenum Press, New York, pp 89–96.
Lewin A, Laufer N, Rabinowitz R and Schenker JG (1986) Ultrasonographi-cally guided oocyte recovery for in vitro fertilization: an improved method. J In Vitro Fertil Embryo Transfer 3,370–373.
Malpaux B, Thiery JC and Chemineau P (1999) Melatonin and the seasonal control of reproduction. Reprod Nutr Dev 39,355–366.
Mathers CD and Harris RS (1983) Seasonal distribution of births in Australia. Int J Epidemiol 12,326–331.
Odegard O (1977) Season of birth in the population of Norway, with particular reference to the September birth maximum. Br J Psychiatr 131,334–339. Ossenbühn S (1998) Exogenous influences on human fertility: fluctuations in
sperm parameters and results of in-vitro fertilization coincide with concep-tions in the normal population. Hum Reprod 13,2165–2171.
Politoff L, Birkhaeuser M, Almendral A and Zorn A (1989) New data confirm-ing a circannual rhythm in spermatogenesis. Fertil Steril 52,486–489. Rameshkumar K, Thomas JA and Mohammed A (1992) Atmospheric
tempera-ture and anovulation in south Indian women with primary infertility. Ind J Med Res 96,27–28.
Reinberg A, Smolensky MH, Halleck M, Smith KD and Steinberger E (1988) Annual variation in semen charcteristics and plasma hormone levels in men undergoing vasectomy. Fertil Steril 49,309–315.
Roenneberg T and Aschoff J (1990a) Annual rhythm of human reproduction. I. Biology, sociology or both? J Biol Rhythm 5,195–216.
Roenneberg T and Aschoff J (1990b) Annual rhythm of human reproduction. II. Environmental correlations. J Biol Rhythm 5,217–239.
Rojansky N, Benshuhan A, Meirsdorf S, Lewin A, Laufer N and Safran A (2000) Seasonal variability in fertilization and embryo quality rates in women undergoing IVF. Fertil Steril 74,476–481.
Rojansky N, Brezinsky A and Schenker JG (1992) Seasonality in human repro-duction: an update. Hum Reprod 7,735–745.
Rosenberg HM (1966) National Center of Health Statistics. Seasonal varia-tion of births: United States 1933–63. Vital and Health Statistics. Series 21, No. 9 (PHS publication no. 1000). US Government Printing Office, Washington, DC.
Saint Pol P, Beuscart R, Leroy-Martin B, Hermand E and Jablonski W (1989) Circannual rhythms of sperm parameters of fertile men. Fertil Steril 51,1030–1033.
Smith DM, Conaway CH and Kerber WT (1978) Influences of season and age on maturation in vitro of rhesus monkey oocytes. J Reprod Fertil 54,91–97. Stolwijk AM, Reuvers MJCM, Hamilton CJCM, Jongbloet PH, Hollanders JMG
and Zeilhuis GA (1994) Seasonality in the results of in-vitro fertilization. Hum Reprod 2300–2305.
Timonen S, Franzas B and Wichmann K (1964) Photosensitivity of the human pituitary. Ann Chir Gynecol Fenn 53,165–170.
Tjoa WS, Smolensky MH, Hsi BP, Steinberger E and Smith KD (1982) Cir-cannual rhythm in human sperm count revealed by serially independent sampling. Fertil Steril 38,454–459.
Udry JR and Morris NM (1967) Seasonality of coitus and seasonality of birth. Demography 4,673–679.
Weigert M, Feichtinger W, Kulin S, Kaali SG, Dorau P and Bauer P (2001) Seasonal influences on in vitro fertilization and embryo transfer. J Assist Reprod Genet 18,598–602.
Wood C, McMaster R, Rennie G, Trounson A and Leeton J (1985) Factors influencing pregnancy rate following an embryo transfer. Fertil Steril 43,245–250.
Wrigley EA and Schofield R (1981) The population history of England, 1541–1871: a reconstitution. Harvard University Press, Cambridge, UK. Yie SM, Brown GM, Liu GY, Collins JA, Daya S, Hughes EG, Foster WG and
Younglai EV (1995) Melatonin and steroids in human pre-ovulatory follicu-lar fluid: seasonal variations and granulosa cell steroid production. Hum Reprod 10,50–55.
Submitted on February 21, 2005; resubmitted on May 22, 2005; accepted on May 27, 2005
